Re: Looking for a case: counting

--- takatunu <takatunu@...> wrote:
> You can say it is undercover ad-verb wearing the
> civilian noun case tag.
> However I learned in school that it is a "complément
> circonstanciel" (de
> temps, de moyen, de manière, etc.) "Circumstance" is
> tagged with a
> "locative" case or clitic in many languages (Hebrew,
> Japanese, etc.) A
> circumstance may determine the verb (hence "adverb")
> but also the clause
> itself. Linguists then say that it determines the VP
> regardless that it
> sometimes really determines an object.
In my feeling, time and space complements should be
considered differently from manner, aim, means etc. Of
course, I also learned at school that these are all
'complements circonstanciels'. But sometimes it's
worth reconsidering school definitions, especially
when many linguists heavily point out that scholar
grammatical definitions have a strong tendency to mix
up syntactic and semantic notions. The usual
definition of an adverb is a very good example.
Definitions of verbs too. Etc.

> Regarding aspect: "Iterative" and other aspects are
> sometimes part of the
> semantic definition of a word such as the verb "to
> repeat" or else at a
> syntactic level like in "many times." I can
> understand that it is
> interesting to point out the first in order to
> differentiate verbs or
> expressions from each other, but I fail to
> understand why some learned
> authors christen extraordinary simple locutions like
> "many times" with big
> names such as "iterative". If so, there should be an
> aspect for "many times
> in a very short time in long sequences" and a
> special "causative" expressed
> in English with the words "for the reason that" or a
> partial causative for
> the Japanese "no sei de". I can make a different
> case or aspect especially
> for any and all words of a language.
That's very interesting. I think that the notion of
process quantifiers is useful because it is very often
used, one way or another depending of the language. It
looks very natural to differenciate between an action
that happens once, an action that happens cyclically,
and an action which is repeated in a +/- messy way.
This notion can apply to many verbal concepts, that's
why it looks rather general. It's not just fancying,
it exists in some languages.

But if you consider the original example, 'to throw
the ball many times', then you must admit that it is
not only iterative and irregular (the ball was thrown
several times, not just once; and it was not thrown
cyclically, but an irregular way, the time intervals
being different and somehow at random), but it is also
quantified as 'many' times; so it is not the same
thing as 'several' times, or 'a few' times, or 'hugely
many times'.

So this makes already 3 different concepts applying to
the verbal expression 'to throw the ball'. Is it
possible to invent a verbal mark that would combine
these 3 concepts ? It probably would be, but the
verbal form would then be rather longer and less
convenient to use for the locutor. And if you had to
add yet more concepts (and they certainly are more of
them which should be handled), then it would get less
and less handy. So the problem is: where shall we
stop, and decide that from now on, we will use
separate words, like 'many times', rather than verbal
marks ? It depends of the purpose of the conlang: if
we want a language that can reasonably be spoken by
real people, we surely have to stop rather early. But
if the aim is to come to a formalization that a
computer can handle, so we can go rather far.

To me the biggest problem is not that we have to add
and add more marks on the verb, its is to decide in
which way these different marks should combine. It's
not just 'Verb + a + b + c', it can be 'Verb + a + b',
then adding a c-mark modifying the modifier a, but not
the modifier b (neither the whole complex) for ex.,
thus sthg like: 'Verb + c(a) + b'

In our example, we could try so:

I throw-PAST-ITE-IRR the ball
(PAST being Past, ITE being Iterative, IRR meaning
Irregularly)

Then if we consider that Iterative itself means : more
than once (more than one time), we should add the
concept 'many' to this concept 'more', coming to
something like: much more than one time, a great deal
more than one time. But this adding doesn't modify in
the least the other elements, like PAST or IRR.

We also could imagine to modify the components PAST or
IRR, like in:

I throw-very(PAST)-ITE-IRR the ball
(I threw the ball several times, in an irregular way,
in an ancient time, = very long ago)

I throw-PAST-ITE-very(IRR) the ball
(I threw the ball several times, in a very irregular
way)

Then you can imagine adding modalities, like:

I throw-PAST-ASS(very(ITE))-IRR the ball
(I threw the ball many times in an irregular way, and
I insist on my assuming that it was MANY times)

or:

I throw-PAST-very(ITE)-DBT(IRR) the ball
(I threw the ball many times, but I doubt it was in an
irregular way)

(Such nuances can be expressed in natlangs by the tone
used by the speaker, or by typography as I just did
for MANY in capitals).

Then we could be more precise regarding the concepts
of 'much', 'many' and so on. For ex. by associating a
scalar value, let's say from 0 to 4, 2 being average
(in that case we should precise which scale do we use,
and what is the value on that scale). Etc.

You're right, this is a no-end process. Surely if we
come to that point, nobody will be able to speak such
a language any more. But a computer could handle it,
provided everything is defined very precisely and
logically. That's what interesting me...



> Don't big words "dissolve" real words into illusive
> abstract tags to show
> that semantics at a word level is unecessary to
> understand language? Why--If
> there is no "many times" anymore but only an
> abstract "iterative aspect
> expressed with an NP", then why bother with the true
> meaning of the words
> "many" and "times" and how they are expressed in
> different languages?
>
But the general concept of 'many' (like its cousin
'much') certainly is a fundamental one. 'Times' looks
more complex, as it can be used in several meanings.
More analysing in perspective...

P.S. Regarding complements like 'with a hammer', which
I repeat is an entirely different concept, I of course
agree that there may be, in some nat- or con-langs,
cases where the verb and the complement aggregate
together (some talk of 'coalescence'), meaning that
the noun concept 'hammer', in its instrumental use,
can become a part of a new verb meaning 'to hit with a
hammer' for ex (? to hammerhit). There are many
examples of that sort, like in French 'lapider' (to
throw stones at somebody - the noun concept 'stones'
being conceptually included in the verb 'lapider',
lapis meaning stone anyway. Interesting that 'lapider'
can NOT be understood like 'to do some effective
action on stones', but 'to do some action on a patient
using stones'. So it is not only the concept of
'stone' that is included in 'lapider', but also the
semantic role of the stone(s) in the action. Also,
'lapider' is usually not throwing just one stone, but
SEVERAL ones. All this is already conceptually
included in the verb 'lapider', even if a superficial
analyse only reveals us the concept of 'stone'
(lapid-er; no other marks visible).



=====
Philippe Caquant

"Le langage est source de malentendus."
(Antoine de Saint-Exupery)

__________________________________
Do you Yahoo!?
Yahoo! Finance: Get your refund fast by filing online.
http://taxes.yahoo.com/filing.html